Proceedings of the 50th Hawaii International Conference on System Sciences | 2017 When humans using the IT artifact becomes IT using the human artifact Dionysios S. Demetis Allen S. Lee Hull University Business School Virginia Commonwealth University [email protected] [email protected] Abstract essay about the concept of the IT artifact per se, we Following Lee & Demetis [20] who showed how do make the argument that the nebulous character of systems theorizing can be conducted on the basis of a that concept is due to a much larger (though subtle) few systems principles, in this paper, we apply these phenomenon at play: the transition of technology principles to theorize about the systemic character of from artifact to system. In fact, the contextual technology and investigate the role-reversal in the richness that has been added to the concept of the “IT relationship between humans and technology. By artifact” after its first use by March and Smith [27], applying systems-theoretical requirements outlined can be re-interpreted as a recognition of such a by Lee & Demetis, we examine conditions for the transition. Inspired by the post-humanist tradition that systemic character of technology and, based on our reflects on the boundaries between humans and theoretical discussion, we argue that humans can technology, we take a different approach and develop now be considered artifacts shaped and used by the a systems theoretical description of the transition (system of) technology rather than vice versa. We from artifact to system. We argue that people are argue that the role-reversal has considerable becoming agents of the (system of) technology. implications for the field of information systems that Over time, technology has penetrated society to has thus far focused only on the use of the IT artifact such a degree that even basic functions now seem by humans. We illustrate these ideas with empirical almost inconceivable without technology. Indeed, the material from a well known case from the financial level of dependence of society on technology has markets: the collapse (“Flash Crash”) of the Dow become so deep that – in a large number of fields – Jones Industrial Average. there are now no manual fallback plans in cases of technological failure. By and large, even when technology fails, we tend to rely on more technology 1. Introduction for rectifying the problems of technological use. Also, the rising trend of technologized decision The field of Information Systems (IS) rests making that has taken certain fields by storm is even largely on examining the contextual use of more alarming. In the foreign exchange markets for technology within social (sub)-systems and example, 85 percent of all trading is conducted by organizations. In such a context, the relationship algorithms alone, i.e., without any human between the social and the technical has always been intervention; this led the scholars that investigated the of special interest to IS researchers. Ultimately, this phenomenon to call this the “Rise of the Machines” interest is applied to the interactions between humans [10]. In the UK, the “ultra-high-speed version of and information technology, and at the center of algorithmic trading, high frequency trading, is attention, one can often find the concept of the IT estimated to account for over 77% of transactions in artifact. From considering the IT artifact as an the UK market” [40, p.5]. ensemble of hardware and software [27] to bundles A skeptic of our position who might seek to argue of material and cultural properties that are against the trend of technologized decision making, recognizable and emerge from ongoing might pose this question: Is it not the case that the socioeconomic practices [34], or even to designers of algorithms are humans? And if so, then sociotechnical assemblages [38], one thing is clear: couldn’t someone consider the role of algorithms the concept of the IT artifact has changed (and of technology at large) as an extended substantially over the years. In fact, the ontological application of human decisions? dimensions upon which the “IT artifact” has come to Our challenge is to convince the reader to the be considered have shifted so much that Steven contrary. For this purpose, we render this issue Alter’s suggestion was to “retire” that concept through a few intertwined questions: How does altogether from the lexicon of IS scholarly debate as technology subvert and subdue human decisions? it has outlived its usefulness [1]. While this is not an 1 URI: http://hdl.handle.net/10125/41855 ISBN: 978-0-9981331-0-2 CC-BY-NC-ND 5747 What characteristics can be identified (with the help bots and websites that are indexed must be of systems theorizing) for this new role that unsupervised. But while the millions of pre-indexed technology has assumed and how does this constitute search results give the illusion of choice, almost 90% the emergence of a system of technology? Even more of humans don’t get past the top ten [18]. The whole crucially, how is it that humans become “artifacts” process feels like a “search on the Internet” but it is being shaped and used by technology in this – actually a restricted human search of a technological seemingly counterintuitive – role reversal? pre-search of the Internet: the “search of a search” As we will see through our example of the Flash that constitutes a 2nd order phenomenon. Hence, this Crash regarding the Dow Jones Industrial Average, is a case of a human reacting to technological stimuli the role of technology leads us to consider a (i.e., an individual person reacting to the seemingly radical idea at first – but one that we algorithmically generated search results intended to believe is an accurate reflection of how technology steer the person’s behavior) rather than a technology shapes social systems and subjects humans to forces reacting to human stimuli (i.e, a neutral search that cast them out to the environment, outside of what algorithm providing objective results to best serve a has become technologized decision making; instead human using the technology). of the IT artifact being shaped and used by humans, Another well-known example comes from humans can actually be considered as “artifacts” Amazon. The vast majority of prices are defined by being shaped and used by machines. In this view, it is algorithms in so far as Amazon vendors “use humans that must react to technological stimuli rather algorithmic pricing to ensure that they can than technology that must react to human stimuli automatically change their product prices based on a where, furthermore, the technological stimuli are competitor” [39], with the result that vendors are emergent and not pre-designed (or pre-programmed) being forced to engage in this practice for fear of in any way. This also assumes that while the losing out to the competition. Meanwhile, the controllability of technology can be achieved at a algorithmic interactions between vendors carry the micro-scale (where one could assert that the link possibility of developing unpredictable between designers and – control of – artifacts is consequences. Such algorithmic pricing on Amazon strict), at a macro-scale, technology exhibits can be found in the example of the book entitled The emergent non-linear phenomena that render human Making of a Fly by evolutionary biologist Peter controllability irrelevant [4, 14]. This creates a Lawrence. This book came to be priced at significant circularity wherein the systemic role of $23,698,655.93 (plus $3.99 shipping) as two sellers technology demands a higher degree of autonomy to were using algorithms to adjust the price of the book be granted in the design of technological artifacts, in response to one another. It took 10 days for where this then leads to a deeper systemic function of humans to notice and intervene to bring back the technology that leads to more demands for autonomy. prices to normal levels [43]; ironically, “normal Humans increasingly find themselves in the levels” merely indicated a temporary human decision environment, outside of these dynamics. that would allow the continuation of algorithmic pricing. 2. Examples Similar examples where human decisions and human reactions find themselves outside the Indeed, one can find a wealth of examples where boundary of technologized decision making can technological autonomy has developed into a system come from any domain. In law for instance, legal that takes over important decisions – and humans analysts are being replaced in complex cases by find themselves outside, i.e., cast out to the software that analyses thousands of legal documents; environment, outside of these decisions. In such proprietary e-discovery algorithms of software examples, human agency – “acting on behalf of companies pre-structure the defense/prosecution of a another, or providing a particular service” [16] is case by effectively pre-selecting an extremely limited being replaced by technologized agency. In subset of documents that will then be looked at and categorizing the World Wide Web for instance, presented in court by humans [28]. By and large, it Yahoo (to bring up one example) uses proprietary would be fair to say that in such cases, humans algorithmic robots (known as bots) to create a become the tool through which computerized searchable database that then ranks users’ search decisions are voiced. In autonomous driving, results based on their search queries. The structuring Google’s self-driving car assumes full control of the